In the prior art, during video playback the relevant information is usually pushed by the following two ways.
First, during video playback a matching picture area (a TV station logo or program icon) may be extracted from the played video frame. The information present in the matching picture area may then be identified using image recognition technology, and the identified information may be matched and retrieved in a database or network. Finally the retrieved result information may be sent as the push information to the user and displayed on a dedicated display area.
Second, during video playback the currently played video frame may be divided into multiple foci, and the image recognition technology may be employed to identify the corresponding image content (e.g., the image may be recognized as a product) of each focus, and then the push information corresponding to the image content may be searched and displayed on a dedicated display area.
Both the above two information push methods require the partitioning of image frames (either extraction of the matching picture area or division of foci) and image recognition technology. Thus, the image processing process is complex and the amount of data computation is high.